The massive growth of sensor devices has assisted the Internet of Things (IoT) technology to reach a greater extent. However, the IoT devices are associated with various limitations, as the attached sensor devices are prone to limited battery lifetime. Though clustering seems to be the outstanding solution, it suffers from various disadvantages including unbalanced energy consumption and high packet loss ratio. In order to overcome these issues, we propose a fog-assisted Multi-Verse Optimization algorithm (MVO) for dynamic clustering of IoT devices (DCMI). The MVO algorithm is a meta-heuristic nature-inspired optimization algorithm possessing powerful operators, which makes the dynamic partitional clustering strategy more efficient. The MVO algorithm in the proposed DCMI approach employs inner and outer fitness function to produce more stable and optimal number of clusters. Moreover, the fog nodes helps the IoT network in transmitting the data with minimal latency. Thus, DCMI approach enhances the network lifetime and minimizes the packet loss ratio. The proposed DCMI approach is simulated and the results are compared with the various existing optimization based clustering approaches. The outcome analysis prove that the DCMI approach outperforms the existing approaches in terms of network lifetime, latency, packet loss ratio, throughput, packet delivery ratio, and energy consumption.